Ineffable Intelligence, the London-based AI startup launched this year by former Google DeepMind researcher David Silver, has brought on six high-profile 'cofounders,' the company confirmed to Fortune.
Four of the new hires — Chris Apps, Wojciech Czarnecki, Lasse Espeholt, and Junhyuk Oh — are former Google DeepMind colleagues of Silver's. Apps, Czarnecki, and Oh worked with Silver on DeepMind's 2019 AlphaStar breakthrough, the system that beat top human players at StarCraft II. Espeholt helped build Google's MetNet weather-prediction models.
The fifth hire, Alexandre Laterre, was formerly head of research at InstaDeep, the London AI firm acquired by biotech company BioNTech in 2023. The sixth, Heather Gorham, was a partner at venture firm Flying Fish, one of Ineffable's earliest investors.
In April, Ineffable was valued at $5 billion after closing a $1.1 billion seed round — the largest seed investment ever raised in Europe. The hiring spree adds to a wave of high-profile researchers leaving Google DeepMind for rival AI companies, feeding a perception that the unit, once first to make artificial general intelligence its stated mission, is now struggling to retain talent.
Companies House records show Ineffable was incorporated in November 2025 by George 'Geordie' Rose, a Canadian serial entrepreneur. Silver did not become a company director until January 2026, after formally leaving DeepMind. Rose lost his status as a 'person with significant control' of Ineffable's shares one month later and stepped down as director in July. A company spokesperson said Rose helped Silver get the company running before moving into an advisory role, but that Silver does not consider him a cofounder.
Ineffable is one of the so-called 'neolabs' — a new generation of frontier AI firms launched in the past two years, many founded by veterans of Google DeepMind, OpenAI and Anthropic. The group also includes Thinking Machines Labs, cofounded by former OpenAI CTO Mira Murati; Safe Superintelligence, cofounded by former OpenAI chief scientist Ilya Sutskever; and Sakana AI, Recursive Superintelligence, Core Automation and Adaption Labs.
Ineffable's stated mission is building AI superintelligence through reinforcement learning, the trial-and-error training method with which Silver is closely associated. Oh, who will run reinforcement learning at Ineffable, was the reinforcement-learning lead on AlphaStar and, in 2025, was among the first researchers to use AI to discover new reinforcement-learning algorithms. Czarnecki, who focused on multi-agent research at DeepMind, will oversee Ineffable's science team.
The numbers come first, and here they tell a clear story: $1.1 billion in private seed capital, a $5 billion valuation, and six senior researchers walking away from one of the world's best-funded corporate labs to join a startup barely a year old. No ministry drafted this roster and no subsidy underwrote the round.
Capital rewards clear rules and cleared incentives — and venture investors, not Alphabet's org chart, are now the ones deciding where the top reinforcement-learning talent in the world goes to work. For an industry Washington and Brussels are both eager to regulate, that is the more instructive headline.
